To perform a verification on an audio, it must meet the requirements of the verification model.

If the model requirements are not met, an error will be returned.

​
Chunk requirements

Authenticity verifications are divided into separate chunks, providing granular results. Each chunk must meet certain requirements to receive a score.

Based on their fulfillment of the requirements, the chunk types are:

  • insufficientVoice — not enough voice was detected.
  • excessiveNoise — too much noise was detected.
  • validVoice — the requirements are met and a score is provided.

If none of the chunks contains valid voice, the resulting verification will not contain a score.

OverviewScores